GH-44953: [R] Add R bindings for new compute functions

[amoeba]

Rationale for this change

Adds R bindings to newly added hyperbolic trig and Expm1 compute functions. See #44953. Bindings prefer calling _checked variants of compute functions where applicable which follows how existing bindings have been implemented.

What changes are included in this PR?

• R bindings for hyperbolic trig compute functions
• R bindings for Expm1(Exponent) compute function
• Adds missing dplyr binding for base::atan
• Adds missing documentation for atan (not directly related to this PR)
• Updates to NEWS.md

Are these changes tested?

Yes.

Are there any user-facing changes?

New compute bindings in R.

• GitHub Issue: [R] Add R bindings for new compute functions #44953

[amoeba]

This PR branch is currently based on #44630 and I'll rebase and force-push once that's merged. I still need to:

• Add the Expm1(Exponent) binding and
• Verify both sets of bindings are tested adequately

[jonkeane]

Thanks for this! One minor comment and one question

[conbench-apache-arrow]

After merging your PR, Conbench analyzed the 4 benchmarking runs that have been run so far on merge-commit 8b0ebb9.

There were no benchmark performance regressions. 🎉

The full Conbench report has more details.

[jonkeane]

Hmmm, coming back and looking at this again, do we need to skip based on version here? These tests were added at the same time the functions were, so won't be run against older versions at all — right? Were they failing for some other reason? I can clean this up in #46294 but maybe I'm missing some reason they are important here?

[thisisnic]

I'm a bit unsure what we have running on our extensive CI these days, but if we don't have anywhere where we run current R arrow combined with older C++ arrow, then yep, we shouldn't need this.

[jonkeane]

We do test across versions of libarrow (maybe that failed when you were looking at this? It doesn't look like it from the commit history, but maybe I'm missing it), though looking at the discussion in #43623 it looks like no one is interested in maintaining + teaching folks about libarrow cross compatibility such that it's worth it to keep that job + the maintenance cost of that.

I'll try removing / reverting in my other PR and see though